Computational protocol: Balancing selection shapes density-dependent foraging behavior

Similar protocols

Protocol publication

[…] The MY14-CX12311 recombinant inbred lines were generated by crossing MY14 males to CX12311 hermaphrodites and CX12311 males to MY14 hermaphrodites, to ensure the mitochondrial DNA from both strains were equally represented in the RILs. 94 F2 were individually picked to plates and inbred through self-fertilization for 10 generations. RIL genotyping was conducted by low-coverage shotgun sequencing. Genomic DNA was fragmented and attached to sequencing adapters with a Nextera DNA Library Prep Kit (Illumina, San Diego, USA). Samples were pooled and sequenced on an Illumina HiSeq 2000. Sequencing reads from each strain were mapped to the WS235 release of the C. elegans genome using bwa to create bamfiles for further analysis. The set of MY14/N2 SNVs identified in the Million Mutation project were used for genotyping purposes. Each genetic variant was genotyped in each strain. Due to the low coverage, the majority of SNVs were not genotyped. To improve the data coverage, we grouped 200 neighboring SNV genotypes together to create a consensus genotype for 540 bins (either N2, MY14 or heterozygous). These genotypes were used for QTL mapping. [...] The pheromone response index was used as the phenotype in combination with the 540 genotype bins from above. R/qtl was used to perform a one-dimensional scan using marker regression on all 540 markers. The significance threshold was determined using 1000 permutation tests. The effect-size of the roam-1 locus was estimated using the fitqtl function with a single QTL. The peak of the roam-1 locus (Chromosome V: 16451686-16579457) was used as an additive and interactive covariate for additional one-dimensional scans, assuming a normal model. The significance threshold for these two tests was also determined using 1000 permutation tests. [...] Calcium imaging experiments were performed and analyzed as described. Briefly, young adult animals were placed into custom-made 3×3 mm microfluidic polydimethylsiloxane devices that permit rapid changes in stimulus solution. Each device contains two arenas, allowing for simultaneous imaging of two genotypes with approximately ten animals each. Animals were transferred to the arenas in S-Basal buffer and paralyzed for 80-100 minutes in 1 mM (−)-tetramisole hydrochloride. Experiments consisted of four 10 s pulses of stimulus separated by 30 seconds of buffer, with 60 additional seconds between stimulus types. Tiff stacks were acquired at 10 frames/second at 5x magnification (Hamamatsu Orca Flash 4 sCMOS), with 10 ms pulsed illumination every 100 ms (Sola, Lumencor; 470/40 nm excitation).Fluorescence levels were analyzed using a custom ImageJ script that integrates and background-subtracts fluorescence levels of the ASH cell body (4×4 pixel ROI). Using MATLAB, the calcium responses were normalized for each stimulus type by dividing fluorescence levels by the baseline fluorescence, defined as the average fluorescence of the 10 s preceding the first pulse of the stimulus. Each experiment was performed a total of four times over two separate days. Animals of a given strain were pooled together to calculate population mean and standard error (N2 srx-43 allele: 23 animals; MY14 srx-43 allele: 30 animals; array negative control: 19 animals). Experiments were conducted on two days.For GFP expression studies, live adult animals were mounted on 2% agarose pads containing 5 mM sodium azide. Images were collected with a 100X objective on a Zeiss Axio Imager.Z1 Apotome microscope with a Zeiss AxioCam MRm CCD camera. For daf-7 reporter studies, expression was quantified 16-24 hours after L4 animals were placed on exploration assay plates. Images were processed in Metamorph and ImageJ to generate a maximum intensity Z-projection. Reporter values were assessed as the mean gray value for a 16-pixel-radius circle centered over the cell body minus the mean background intensity. Both ASI neurons were analyzed in each animal; experiments were performed on three days. [...] To create the gene and organism phylogenies, we used SNV data downloaded from the Million Mutation project (http://genome.sfu.ca/mmp/) or the CeNDR resource (http://www.elegansvariation.org). For the CeNDR dataset, MY14 was assumed to be clonal or near-clonal with MY23, as was suggested by RAD sequencing. Software was written in Python using the Biopython module to create a neighbor joining tree. For the roam-1 locus, SNVs on Chromosome V between 16,010,000 and 16,030,000 were used. For the glc-1 locus, SNVs on Chromosome V between 16,181,000 and 16,222,000 were used. All SNVs were used to construct the whole genome strain tree. Number of genetic variants and Tajima's D were calculated on 5 kb bins using vcftools. dN/dS was calculated by counting using custom Python scripts analyzing variants between MY23 and the N2 reference. Phylogenies of srx-43 and closely related genes were performed using protein sequences from Thomas and Robertson, 2008. […]

Pipeline specifications

Software tools BWA, R/qtl
Application WGS analysis
Organisms Caenorhabditis elegans